To the Editor: We noted with interest the findings of William T. Howard, M.D., M.S., et al. (1) and the discussion by Katherine A. Halmi, M.D. (2). They reported that inpatients who had not attained a body mass index of 19 in the hospital did less well in a day program setting. In our longitudinal study (3), we found that 50 such patients were still doing less well 6–10 years later (mean=8 years). There was a significant difference in total bone mineral density (p<0.05) between former inpatients with a body mass index of ≥19 after refeeding (osteoporosis in the femoral head=7%, osteoporosis in the lumbar spine=0%) and those who had a body mass index of <19 after refeeding (24% and 12%, respectively). Scores on the Eating Attitudes Test and the Drive for Thinness subscale of the Eating Disorders Inventory were significantly higher when body mass index was <19, and severe depression was reported only in this group. Outcome on the Morgan Russell criteria was more often in the poor category (22% versus 11%), and reproductive function was more compromised. Thus, failure to reach treatment goals early had long-lasting effects. When one considers the public health costs of osteoporosis and the private health costs (and the dangers) of assisted fertility—not to mention the costs of depression, chronic ill health, ongoing eating disorders, and the effect on the next generation—it can only be hoped that health funding bodies can recognize the false economy of shortening the period of nutritional rehabilitation in anorexia nervosa.